Automation Tools
for Better Code

Automated Software Parallelization for Embedded Multicore, GPUs, DSPs and Vector Processors

With more than 10 years of research and tool development, emmtrix Technologies is your solution provider for efficient software development targeting high-performance embedded multicore systems: Multicore and vector processors, accelerators (e.g. GPUs, DSPs), or any heterogeneous combination of these are fully supported.

Your benefits:

Parallelize your applications onto the various processing elements (software parallelization)
Maintain your functional safety (ISO 26262 / DO 178C)
Convert your code (C++, MATLAB®, Simulink™, GNU Octave and Scilab) efficiently and transparenly to C
Estimate the performance of your code even at an early development stage
Understand data dependencies and event chains of your code (static code analysis)

Our Solutions

Performance Estimation

Early in workflow
Master your timing budgets
Continuous integration support
Estimation methods:
- Static source code
- Simulation
- Profiling on hardware
Intuitive visualization
Continuous performance monitoring during the development

Parallelization

Parallel C code for
- Multi-/Manycore CPUs
- Vector Processors
- GPUs
- DSPs
Interactive workflow
Parallelization with functional safety according to standards like ISO 26262, DO-178C and others

Automated Vectorization

Easy exploitation of parallel vector hardware
Correct-by-construction code generation
Speedup > 10x

Data Dependency Analysis

Identify data dependencies
Verify your specification
Document for (re-)certifcation
Event chain analysis
Data flow analysis

Code Conversion

Conversions:
User-controlled optimizations
Aimed at embedded systems and automatic analysis

Tools & Services

emmtrix Parallel Studio (ePS)

User-configurable automated development of parallel C code for embedded multicore systems, GPUs and FPGAs.

emmtrix Code Generator (eCG)

Embedded C code generation from MATLAB® or Simulink®, Scilab or Xcos or GNU Octave optimized for parallelization.

emmtrix Model Code Generator (eMCG)

Translate graphical Simulink® models (*.slx) into MATLAB® or GNU Octave language files (.m) with emmtrix Model Code Generator.

emmtrix C++ to C Compiler (eCPP2C)

Automatical translation of C++ source code into analyzable C code enables software parallelization of C++ applications with ePS.

ePS Qualification Kit (ePS-QKIT)

Verification tool for emmtrix Parallel Studio according to DO-178C/330, ISO 26262 and comparable functinal safefty standards.

Customization, Consulting and Trainings

Wide range of custom services to address your specific multicore requirements, get going fast and take the right decisions.

Free Webinars on Software Optimization

Modern multi-core systems, DSPs and GPUs are ubiquitous, with vector processing units like the PPU of AURIX TC4x™ adding another type of accelerator. In order to use these platforms efficiently and effectively, companies and software developers have to deal with and understand parallelization/vectorization. Distributing software onto the available processing units increases performance but also complexity. In our webinars you will learn how to successfully parallelize/vectorize code, as well as more about our tools and their benefits.

View webinars

Latest News & Next Event(s)

Data Dependency Analysis

Solution | Jun 01, 2022
Analyze the data dependencies in the source code of your application. Test our new solution now.

Resellers - Success Factors

Interview | May 17, 2022
Part II: Tsunemori Kawahara introduces NeXtream and speaks about success in JP.

Resellers - Success Factors

Interview | May 10, 2022
Part I: Rainer Heim talks about the benefits of selling through a reseller and NeXtream.

Presentation:
June 30, 2022
11:45 am – 12:30 pm
Room: Vienna/Athens

Jun 28 – 30, 2022 | Hybrid Event

SAEC Days 2022

Make sure to participate at our presentation “Software-Parallelisierung unter Berücksichtigung der funktionalen Sicherheit” (“Software Parallelization Considering Functional Safety”) at SAEC Days 2022. In our presentation, we will discuss important properties in the comparison of semantic equivalence, how parallelization should be done, and what a possible workflow might look like.

View presentation details and get your ticket now:

Visit event page